1. State your health outcome of interest. (It could be the one you used for week #2 or another one.) Pick two key behaviors that are important factors leading to your health outcome. Explain the importance of these behaviors either for etiology, prevention, or intervention. (If none of the behaviors in the readings are important for your health outcome, suggest another behavior that is.)
Health Outcome of interest: Recidivism after initial Juvenile Justice Involvement
Key Behaviors: 1. Cumulative traumatic experiences: Many incarcerated youth endure a history of trauma that has been linked to poverty and residence in urban areas with community violence (Kliewer & Lepore, 2015). Experiencing cumulative forms of trauma (e.g. interpersonal violence, sexual abuse, and neglect) is one of the most salient predictors of future risk behaviors (e.g., delinquency). 2. Substance Use: Several cross-sectional studies have identified that justice-involved youth exhibit significant substance use (Dembo et al., 2009) and that substance use is associated to future recidivism (Langan & Levin, 2002; Hanlon et al (1998).
2. Describe how you would study the role of one of the behaviors described for question #1 and your health outcome of interest. Incorporate a social factor (e.g. race/ethnicity, social exclusion, stress) in the study approach.
(1) In order to study the behaviors of interest (substance use and trauma), I’d first investigate group differences by gender (i.e., Male vs Female) for both (a) the prevalence and co-occurrence of substance use and HIV/STI risk behaviors at baseline as well as the interrelationships of acute trauma symptoms, recent substance use, and HIV/STI sexual risk behaviors. More specifically, whether increased levels of acute trauma symptoms are associated with greater recent (past 90 days) substance use and HIV/STI risk.
(2) Next, I’d assess whether gender and race/ethnicity predict patterns of future substance misuse, sexual risk behaviors and recidivism.
To test Aim 1, I’d use basic descriptive statistics and between group analyses (e.g., t-test and chi-square) to describe the differences in the prevalence of trauma exposure and symptoms, as well as the outcomes of interest (e.g., substance use, HIV/STI sexual behavior). I would then use logistic regression to examine correlates and risk factors for substance use and HIV/STI sexual behavior at baseline. To test for group differences by gender, I would conduct stratified analyses and will include interaction terms between gender and covariates of interest (e.g., trauma exposure) as appropriate.
To test Aim 2, I would first examine group differences by gender and race/ethnicity using ANOVA and ANCOVA models (for continuous variables) as well as analyses appropriate for dichotomous outcomes (e.g., Mantel-Haenszel chi-square). Gender will also be examined as a moderator of intervention outcomes over time for subgroups of youth. Specifically, I would include interaction terms between gender and covariates of interest (e.g., trauma symptoms) to determine whether the effect of these variables on primary outcomes significantly differs by gender and race/ethnicity.
3. If key health behaviors (e.g. smoking, exercise, nutritious diet) are strongly influenced by neighborhood, income, and/or education, do we need to continue to study how these behaviors influence health outcomes? Why or why not?
Yes. Income and neighborhood greatly impacts youth. School policies such as the zero tolerance policy are implemented at low income schools increasing the presence of law enforcement and contact with the juvenile justice system compared to more affluent youth where school professionals handle the disciplinary actions. Neighborhoods impact the stress and and presence of law enforcement, placing youth that live in these neighborhoods at greater risk for contact with the juvenile justice system.